The human JAZF1 locus is notoriously associated with type 2 diabetes (T2D). A gene variant defined by a single nucleotide polymorphism (SNP) mapping to intron 1 (rs864745-T) is the most widely reported to associate with T2D and was first identified as a significant T2D risk allele in a meta-analysis of genome-wide association studies (GWAS) conducted in individuals of European descent (Zeggini et al., 2008). Several population-based studies across ethnic groups have successfully replicated the finding (Omori et al., 2009; Rees et al., 2011; Cooke et al., 2012; Langberg et al., 2012; Ng et al., 2013; Chang et al., 2014; Song et al., 2015; Deng et al., 2017; García-Chapa et al., 2017). Other T2D-associated SNPs have also been identified on JAZF1 (Sollis et al., 2022), many of which have additionally been linked to diabetes-relevant metabolic traits such as insulin secretion (Grarup et al., 2008; Boesgaard et al., 2009; Dai et al., 2022), insulin clearance (Goodarzi et al., 2013), plasma lipid levels (An et al., 2009; Dai et al., 2022) and body weight (Grarup et al., 2008; Kobiita et al., 2020; Zeng et al., 2021). Interestingly, all of the reported T2D risk SNPs on the JAZF1 locus fall within a region of intron 1 thought to play a role in regulating JAZF1 expression (Ding et al., 2022). Particularly, the T2D risk allele rs1635852-T (van Hoek et al., 2008) exhibits cis-regulatory activity over JAZF1 in pancreatic islets, resulting in reduced gene expression (Fogarty et al., 2013; Kobiita et al., 2020), while the rs864745-T allele correlates with lower JAZF1 expression in skeletal muscle (Ding et al., 2022). These observations hint at the importance of adequate JAZF1 expression in maintaining a healthy metabolism and preventing metabolic disease. Indeed, JAZF1 is found downregulated in pancreatic beta-cells of diabetic patients (Marselli et al., 2010; Taneera et al., 2012; Kobiita et al., 2020), in liver cells of hepatic steatosis patients (Wei et al., 2018) and in metabolic tissues of diabetes and obesity rodent models (Li et al., 2014; Wei et al., 2018; Kobiita et al., 2020; Sághy et al., 2020; Zhou et al., 2020). Accordingly, Jazf1 KO mice develop insulin resistance that is further exacerbated by high-fat diet feeding (Lee et al., 2022). In addition to genetic factors, there is evidence that the expression level of JAZF1 can be influenced by the metabolic context. Mice under a high-fat diet, mouse pancreatic islet cells cultured in high glucose and primary human islets cultured in a diabetogenic medium all show a decrease in JAZF1 mRNA and protein levels (Ng et al., 2013; Li et al., 2014; Wei et al., 2018; Kobiita et al., 2020).

To shed light on the metabolic function of JAZF1, several groups have investigated its activity in mammalian metabolic organs, including the pancreas, liver, adipose tissue, skeletal muscle and the hypothalamus. In vivo and in vitro studies alike argue for a pivotal role in the maintenance of glucose and lipid homeostasis, where JAZF1 enhances insulin signaling, promotes lipolysis and glucose uptake, and suppresses lipogenesis and gluconeogenesis in insulin-responsive tissues. Overexpression of JAZF1 upregulates the lipolytic enzymes HSL and ATGL and the glucose transporters GLUT1, GLUT2 and GLUT4, downregulates the lipogenic enzymes SREBP1, ACC1, FAS and HMGCR and the gluconeogenic enzymes PEPCK and G6Pase, and leads to increased phosphorylation of AMPK, IR, IRS-1 and Akt as well as increased insulin-stimulated PIP3 production, suggesting improved insulin sensitivity via the IR-PI3K-Akt pathway (Li et al., 2011; Ming et al., 2014b; Jang et al., 2014; Li et al., 2014; Yuan et al., 2015; Wei et al., 2018; Zhou et al., 2020; Ding et al., 2021; Ding et al., 2022). Consequently, JAZF1 overexpressing animals are protected against diet-induced insulin resistance and related metabolic dysfunctions (Jang et al., 2014; Li et al., 2014; Yuan et al., 2015; Wei et al., 2018; Zhou et al., 2020). Furthermore, JAZF1 has been linked to other equally important, metabolically relevant processes, including pancreatic beta cell differentiation and function (Taneera et al., 2012; Kobiita et al., 2020; Park et al., 2021), ribosome biogenesis (Kobiita et al., 2020; Procida et al., 2021) and adipogenesis (Ming et al., 2014a; Jeong et al., 2021).

These valuable functional insights warrant a mechanistic dissection of the biochemical function of JAZF1, which is still lacking. Nevertheless, a few clues could be found in recent publications. JAZF1 was initially identified as a direct interactor of NR2C2, a nuclear hormone receptor and transcription factor with context-dependent activator or repressor function, where JAZF1 inhibited NR2C2 activity on a reporter plasmid without affecting its homodimerization or DNA binding efficiency. It was therefore suggested that JAZF1 was a co-repressor of NR2C2 (Nakajima, 2004). However, most of the literature characterizing JAZF1 at the biochemical level argues for a co-activator function. JAZF1 has been reported to associate with the NuA4/TIP60 transcription co-activator complex in different human cell lines (Piunti et al., 2019; Procida et al., 2021; Alerasool et al., 2022). Using a reliable approach for the purification and characterization of native protein complexes, consisting of tandem affinity purification of the target protein at physiological expression levels followed by proteomic analysis (Dalvai et al., 2015), our group has established JAZF1 as a stable stoichiometric subunit of NuA4/TIP60 (Sudarshan et al., 2022). Moreover, a chromosomal translocation frequently occurring in a type of sarcoma which merges JAZF1 and SUZ12, a subunit of the co-repressor methyltransferase complex PRC2, mistargets the acetyltransferase activity of NuA4/TIP60 to PRC2-occupied loci, leading to their upregulation (Sudarshan et al., 2022; Tavares et al., 2022). In addition, JAZF1 seems to positively regulate acetylation of the histone H2A.Z at intronic enhancers, a chromatin modification that is specifically deposited by Tip60 (Procida et al., 2021). Remarkably, JAZF1 was detected as a top hit in a dCas9-based screen for transcription activators and co-activators in human cells (Alerasool et al., 2022). While we detected NR2C2 in our tandem affinity based-proteomic analysis of JAZF1 interactome, it was clearly sub-stoichiometric relative to NuA4/TIP60 subunits with similar molecular weights (Sudarshan et al., 2022). This suggests that the complex might be acting as a co-activator of NR2C2, with JAZF1 mediating the interaction. In fact, genome-wide NR2C2 binding sites in different human cell lines are enriched in RNA processing and protein translation genes (O'Geen et al., 2010), similar to JAZF1 (Kobiita et al., 2020; Tavares et al., 2022) and other NuA4/TIP60 subunits (Jacquet et al., 2016). The same study also found that NR2C2 exhibited cell type-specific binding at genes involved in lipid and carbohydrate metabolism, and that most of NR2C2 target loci are active genes (O'Geen et al., 2010). Overall, the current literature is in favor of a model whereby JAZF1 carries out its metabolic functions as an integral component of the acetyltransferase complex NuA4/TIP60, with possible cooperation with the transcription factor NR2C2 (Figure 2).

In S. cerevisiae, the transcription factor Sfp1 bears significant homology with mammalian JAZF1 (Kobiita et al., 2020; Sudarshan et al., 2022) (Supplementary Table S1). Sfp1 is best known for its critical role in modulating yeast growth, where it tightly controls the growth rate and positively regulates cell size at division (Jorgensen et al., 2002; Cipollina et al., 2008; Fazio et al., 2008). Mechanistically, Sfp1 achieves this through its function as a major transcriptional activator of protein biosynthesis genes, including ribosome biogenesis, ribosomal protein and other translation-related genes (Jorgensen et al., 2002; Fingerman et al., 2003; Cipollina et al., 2008; Fazio et al., 2008; Albert et al., 2019; Zencir et al., 2020). Crucially, there is mounting evidence of a potential interaction between Sfp1 and yeast NuA4. An interactome analysis found that a subunit of NuA4, Tra1, associates with Sfp1 (Lempiäinen et al., 2009), and an acetylome analysis detected NuA4-dependent acetylation of Sfp1 (Downey et al., 2015). However, unlike JAZF1, Sfp1 does not seem to be a stable subunit of the NuA4 complex, considering the low recovery of NuA4 subunits in the interactome analysis (Lempiäinen et al., 2009). Thus, the biochemical nature of interaction appears to be different in yeast. Nonetheless, as in human, the yeast NuA4 complex occupies ribosomal protein and ribosome biogenesis genes (Rossetto et al., 2014), suggesting that Sfp1 and NuA4 co-regulate expression at these loci and arguing for some functional conservation relative to higher eukaryotes.

Sfp1 activity is highly sensitive to metabolic cues. For instance, glucose stimulates translocation of Sfp1 from the cytoplasm into the nucleus and leads to Sfp1-mediated gene activation (Marion et al., 2004; Cipollina et al., 2008; Albert et al., 2019). Interestingly, this nuclear-cytoplasmic shuttling of Sfp1, the main mechanism for regulation of its activity, is governed by the kinase TORC1, a major metabolic signaling hub (Marion et al., 2004; Lempiäinen et al., 2009; Zencir et al., 2020). TORC1 directly phosphorylates Sfp1, resulting in its localization to the nucleus and subsequent binding and upregulation of target genes (Lempiäinen et al., 2009). This raises the question of whether JAZF1 could similarly be modified and modulated by TORC1 in mammals (Figure 2A).

Aside from protein synthesis, Sfp1 has been reported to positively regulate glycolysis via upregulation of the glycolytic gene HXK2 (Cipollina et al., 2008), implying a role in glucose homeostasis, and to negatively regulate lipid droplet formation via downregulation of the lipogenic gene DGA1 (Teixeira et al., 2021), suggesting a role in lipid homeostasis, all reminiscent of functional data on JAZF1. Curiously, Sfp1 inhibits TORC1-mediated phosphorylation of Sch9, the closest yeast homolog to the mammalian Akt proteins (Supplementary Table S1). This seems to be a negative feedback loop, as the two act in parallel pathways downstream of TORC1 to induce ribosome biogenesis and ribosomal protein gene activation (Jorgensen et al., 2002; Jorgensen et al., 2004; Lempiäinen et al., 2009).

Both mammalian and yeast NuA4/TIP60 have been linked to the same metabolic processes described for JAZF1 and Sfp1, either via the catalytic acetyltransferase subunit or other proteins in the complex (Supplementary Figure S1). A CRISPR knock-out screen in mouse pancreatic beta cells identified seven NuA4/TIP60 subunits (Tip60, ING3, DMAP1, p400, BRD8, BAF53A and MRGBP) as positive regulators of insulin secretion, involving the complex in beta cell function and arguing for a potential role in diabetes (Fang et al., 2019). Moreover, mammalian Tip60 and BRD8 are essential for adipogenesis and together with p400 are recruited as co-activators by the key adipogenic transcription factor PPARγ (van Beekum et al., 2008; Grönniger et al., 2010; Couture et al., 2012). Additionally, enzymatic activity of PCK1, the rate-limiting enzyme in gluconeogenesis, is positively controlled by Esa1/Tip60-mediated acetylation in yeast and human, supporting a role in glucose homeostasis (Lin et al., 2009). In parallel, disruption of yeast NuA4 leads to increased glycogen biosynthesis through direct regulation of protein kinase A (Filteau et al., 2015; Walden et al., 2020). Further, implication of the NuA4/TIP60 complex in lipid homeostasis is clear, albeit somewhat controversial. In S. cerevisiae, NuA4 subunits Esa1, Eaf1 and Eaf7 promote fatty acid biosynthesis and lipid droplet formation by enhancing activity of the fatty acid biosynthesis enzyme ACC1 (Dacquay et al., 2017; Rollins et al., 2017; Pham et al., 2022). In mammals, there is evidence for Tip60 functioning in the opposing processes of lipogenesis and lipolysis. Tip60-mediated Lipin-1 acetylation is an activating modification which triggers Lipin-1 translocation to the ER where it catalyzes a critical step in triacylglyceride synthesis. This process is likely conserved in yeast (Li et al., 2018). Yet, Tip60-mediated acetylation of Pacer, an important effector in the process of autophagy, negatively regulates lipid droplet accumulation (Cheng et al., 2019). In parallel, during glucose deprivation, Tip60 promotes lipolysis of lipid droplets through direct acetylation of choline kinase (CHK) alpha2 (Liu et al., 2021). Finally, Tip60 suppresses the PI3K-Akt pathway in some types of cancer (Yang et al., 2017; Zhang et al., 2018), while Esa1 suppresses the equivalent pathway by acetylating Sip2 (homolog of the mammalian AMPKβ) which promotes its interaction with Snf1 (homolog of the mammalian AMPKα) and inhibition of Snf1-mediated phosphorylation of Sch9 (homolog of Akt) (Lu et al., 2011) (Supplementary Table S1).